Line tracking performance of a least squares adaptive lattice algorithm

Abstract

The work shows the performance of an algorithm for adaptive lattice structures in line tracking. The algorithm introduces some modifications in the classical LMS procedure, which theoretical background can be viewed in (1). Basically, the algorithm uses an adaptive selection for the two basic parameters 6 and y in an LMS algorithm for all-pole lattice structures. The so-called memory parameter 6 and weighting residual parameter Y are updated at each sample of the input signal and also they exhibit its dependence with the section number along the lattice structure. This work will show up to what degree these parameters can improve the tracking behaviour of the procedure.